discussion Papers How Much Do Perceptions of Corruption Really Tell Us? Claudio Weber Abramo Transparência Brasil

Transcription

1 discussion Papers Discussion Paper May 4, 2007 How Much Do Perceptions of Corruption Really Tell Us? Abstract: Claudio Weber Abramo Transparência Brasil Please cite the corresponding journal article: Regressions and tests performed on data from Transparency International Global Corruption Barometer 2004 survey show that personal or household experience of bribery is not a good predictor of perceptions held about corruption among the general population. In contrast, perceptions about the effects of corruption correlate consistently among themselves. However, no consistent relationship between opinions about general effects and the assessments of the extent with which corruption affects the institutions where presumably corruption is materialized is found. Countries are sharply divided between those above and below the US$ 10,000 GDP per capita line in the relationships between variables concerning corruption. Among richer countries, opinions about institutions explain very well opinions concerning certain effects of corruption, while among poorer countries the explanatory power of institutions for the effects of corruption falls. Furthermore, tests for dependence applied between the variables in the sets of respondents for each of 60 countries also show that, for most of them, it is likely that experience does not explain perceptions. On the other hand, opinions tend to closely follow the trend of other opinions. Additionally, it is found that in the GCB opinions about general effects of corruption are strongly correlated with opinions about other issues, as much as to justify the hypothesis that it would suffice to measure the average opinion of the general public about human rights, violence etc. to accurately infer what would be the average opinion about least petty and grand corruption. The findings reported here challenge the value of perceptions of corruption as indications of the actual incidence of the phenomenon. JEL: D73, H11, K42. Keywords: corruption, perceptions, corruption indicators. Correspondence: Transparência Brasil. Rua Francisco Leitão 339, cj São Paulo (SP), Brazil. Voice: ; Fax: same; crwa@transparencia.org.br; website: Author(s) This work is licensed under a Creative Commons License - Attribution-NonCommercial 2.0 Germany

2 How Much Do Perceptions of Corruption Really Tell Us? Claudio Weber Abramo* Transparência Brasil * The author is especially thankful to Nick Duncan for extensive corrections and suggestions. Daniel Treisman and Eric Uslaner expressed valuable criticism about the strength of the evidence presented here. Jens Andvig, Michael Johnston and Tina Søreide also offered comments. Thanks also to Transparency International, for having made available micro data from its Global Corruption Barometer The usual disclaimers apply.

3 Summary... 3 Introduction... 4 The data and the model... 6 Cross -country analysis... 8 The unusual behaviour of perspectives about the future...19 Intra-country relationships between variables Other opinions Consequences for the interpretation of the CPI Consequences for the interpretation of the DKM index Objections Annex I Questions and dummies Annex II Place selections Indices of tables and figures References Data

4 Summary Perceptions of corruption are often taken as reliable proxies for the actual phenomenon of corruption occurring in countries. Regressions and tests performed on data from the Global Corruption Barometer 2004 (GCB), 1 however, show that personal or household experience of bribery is not in fact a good predictor of perceptions held about corruption among the general population. Countries are sharply divided between those above and below the US$ 10,000 GDP per capita line in the relationships between variables concerning corruption, and especially those involving experience vs. opinions. It is found that the connection between the experience variable and the others mostly remains weak or non-significant in both groups. Controlling for GDP per capita leads to very small and non-significant correlations almost completely across the board. In contrast, perceptions about the effects of corruption correlate consistently among themselves, the exception being the outlook towards the future, which does not appear to be connected to any of the other factors. No consistent relationship between opinions about general effects and the assessments of the extent with which corruption affects the institutions where presumably corruption is materialized is found. Among richer countries, opinions about institutions explain very well opinions concerning certain effects of corruption, while among poorer countries the explanatory power of institutions for the effects of corruption falls, sometimes radically. Furthermore, tests for dependence applied between the variables in the sets of respondents for each of 60 countries (48,232 in all) show that, for most of them, it is likely that experience does not explain perceptions. On the other hand, opinions tend to closely follow the trend of other opinions. For example, in most countries, an opinion that petty corruption is a problem is not significantly more frequent among respondents that have had experience with bribery than otherwise while for most countries the opposite happens concerning the relationship between perceptions. Additionally, it is found that in the GCB opinions about general effects of corruption are strongly correlated with opinions about other issues, as much as to justify the hypothesis that it would suffice to measure the average opinion of the general public about human rights, violence etc. to accurately infer what would be the average opinion about at least petty and grand corruption. 1 A survey on corruption-related issues conducted by Transparency International (TI) encompassing 60-plus countries. 3

5 The lack of sufficient explanatory power of citizen s experience reported in the GCB extends to external opinions such as those systematized in Transparency International s Corruption Perce p- tions Index and other sim ilar indices. 2 The findings reported here challenge the value of perceptions of corruption as indications of the actual incidence of the phenomenon. Also, as the relationship of experience and perceptions, as well as those between perceptions, vary between countries (and substantially vary between rich and poor countries), it is likely that different factors affect the formation of opinions in different environments. This not only makes understanding perceptions countrydependent but also compromises the informative content of rankings of countries based on perceptions of corruption. Introduction The difficulties of directly measuring corruption have led to various alternative methods of estimating the phenomenon from other data. Foremost among them are indices based on opinions about corruption. The most famous of them is the Corruption Perceptions Index, issued yearly by Transparency International. A more ambitious enterprise along the same line has been pursued by Daniel Kaufmann et. al. in the World Bank s Governance Indicators led by that author. The World Economic Forum s Executive Opinion Survey also includes questions about corruption. Every time these indices are announced, they are presented in the press as indices of corruption. More often than not, the perceptions part is forgotten. 3 This leads most lay persons to take such indices as reflecting actual levels of corruption affecting countrie s even if, as pointed out by many authors, 4 the meaning of actual level of corruption is not at all clear. Taking perceptions as indications of actual phenomena by default can become a habit. Thus, after arguing that (for instance) conviction rates are not an adequate indicator for the actual inc i- dence of corruption, but rather, reflect the quality of the judiciary (which in itself is sensible), Lambsdorff (1999) proceeds to state that perceptions are commonly a good indicator of the real 2 It should be borne in mind that the samples basing the CPI and the GCB are different. 3 Press releases from Transparency International itself often fall into the same trap. 4 See e.g. Søreide (2006). 4

6 level of corruption, as if there were some independent estimate of a real level (whatever that means) against which the perceptions could be compared. Given the influence of indices based on perceptions on the public opinion of each particular country and even on informing policies from governments and donor agencies, 5 the question as to whether opinions about corruption actually mean anything beyond themselves is quite crucial. However, testing the matter using only the data leading to those indices is impossible, because they are limited to opinions. In 2004 an ample survey on corruption by Transparency International (the Global Corruption Barometer, subsequently re-issued in 2005 and 2006) allowed for the first time to compare opinions with experiential data collected from the same sample. The present study uses the GCB-2004 to perform extensive statistical testing both on the aggregated country data and the raw data collected in each surveyed country. Several regressions are performed over the aggregated data from countries in order to ascertain whether or not the average opinions expressed hold significant relationships with the average experience with corruption reported by respondents. The result is that perceptions are not good predictions for experiences. On the other hand, perceptions are mostly good predictors (sometimes excellent predictors) of other perceptions, not only related to corruption but also to other, apparently unrelated, matters. It seems that opinions operate in a coherent world. The problem is that such imaginary world of opinions and guesses seems not to hold a close relationship with the world of reality, at least in what regards corruption. The availability of the raw data from each country surveyed by the GCB allows for a more intimate examination of the relationships between experience and opinions within each country. The method used with each set of national data was to test the dependence between variables. A high dependence between the variables representing opinions and the (single) variable representing experience with corruption would lead one to conclude that, more likely than not, there is coherence 5 The Fox administration in Mexico explicitly based its (failed) anti-corruption strategy on the placement of that country in TI s Corruption Perceptions Index. USAID, under the Bush administration, announced that it will use, or is already using, corruption perceptions indices when deciding about whether or not giving aid to countries. IFIs such as the World Bank often play with the idea of fo llowing the same path. 5

7 between them. Lack of dependence would indicate that opinions behave randomly vis à vis exper i- ence and this is what is mostly found. Moreover, the disparities between opinions and experience follow no common pattern from country to country. If a reasonably common pattern were found, then one could uniformly factor out distortions in the opinion-forming processes of people in order to compare countries. The absence of such commonality reinforces the conclusion, already stemming from the analysis of the aggregated data, that the distance between opinions and experiences vary haphazardly from country to country and therefore perceptions-based comparisons between countries cannot be linked to differences in the underlying material conditions. Of course, strictly speaking, the conclusions of this study apply only to the specific survey that led to the Global Corruption Barometer. In order to ascertain whether or not similar disparities between perceptions and experiences obtain in other surveys, it would be necessary that these included at the same time questions about perceptions and questions about experiences. However, the most popular corruption perceptions indices (TI s and the World Bank s) are not even one single survey, being aggregations of widely disparate independent surveys. Some of them might include questions about experiences, but neither Transparency International nor the World Bank Institute ever indicated an interest in adding such data to the analysis of their respective indices. The data and the model The Global Corruption Barometer was a public opinion survey conducted between July and September 2004 by Gallup International on behalf of Transparency International among 52,682 respondents from 64 countries. The survey questionnaire uses in part Gallup s Trust in Instit u- tions survey, which have been done several times in Iberoamerica (14 Latin American countries plus Spain and Portugal), 7 added by one question concerning experience with bribery and numerous questions about perceptions of corruption and other issues. National samples varied from national to urban to metropolitan, the majority of them being urban. Three methods were used: Face-to-face interviews, telephone interviews, in one case (Japan) self-applied questionnaires and in another 6 Details at 7 See the 2002 edition at 6

8 (Norway) web interviews. Samplings were mostly based on demographic quotas, while in a few cases they were probabilistic. Thus, margins of error vary considerably from country to country. The informed overall margin is ±4.4 pp. The questions asked in the survey are presented in Annex I, together with the list of dummy variables we will use, built upon the answers to them. A total of 28 questions were asked. We will classify them into four categories and study the relationships between the variables according to their pertinence to these categories: Group 1. Effects of corruption Petty, Grand, Life, Business, Politics, Perspective. Group 2. Institutions Customs, Education, Judiciary, Health, Police, Parties, Parliament, Civil Registry, Utilities, Taxes, Private Sector, Media, Military, NGOs, Religions. Group 3. Includes just Experience. Experience is not referred to specific Institutions. Group 4. General issues Prices, Poverty, Environment, (Human) Rights, Violence and Jobs. All groups but Group 3 concern opinions. Our attention will be focused on the relationship between experience with bribery and opinions about corruption, be them about Effects or Institutions. Our aim is to find out the extent with which personal/household experience with bribery informs the opinions of people. We will also be interested in ascertaining the relationship between opinions. The analysis will be restricted to 60 of the 64 countries depicted in the GCB (48,232 individual respondents). Three of the countries did not include some or all the questions we are interested in and for one (Kosovo) there are no economic data easily available. For GDP per capita we used the International Monetary Fund s data. We also used Transparency International Corruption Perce p- tions Index and the recent index proposed in Dreher et al. (2004). GDP-PC is used to group countries into two categories, divided by the US$ 10,000 line. There would be 24 countries in the upper tier and 36 in the lower group. Because for almost all variables two countries in the upper tier (Greece and South Korea) behave much closer to the bottom tier than to the top, we grouped then together with the lower-income group. Thus, we ended up with a Top 22 and a Bottom 38 assemblies. Besides testing aggregated data, we performed tests for dependence between variables within each country s set of individual responses. The method used to do this is described in Annex II. 7

9 Cross-country analysis Understanding the phenomenon of corruption is difficult because of its secret character. Not being amenable to direct measurement, corruption is addressed by indirect means, the most prominent being perceptions as measured by Transparency International s Corruption Perceptions Index (CPI), which is built upon a number of different surveys. There are other global indices in existence, such as the World Economic Forum s and, notably, the Control of corruption variable included in the World Bank Institute s KK set of governance indicators. All these are totally 8 or mainly based on opinions of respondents from or in some way related to business, and in a good measure to transnational business. Thus, the representativeness of the samples basing those indic a- tors is very limited. 9 Nevertheless, due to the lack of other measurements, they, and especially the most popular one, the CPI, are taken as depicting countries aggregate levels of corruption the perceptions part being often forgotten. The limitations of the traditional perceptions indices, together with the lack of sufficient guarantees that respondents to those surveys hold intersubjective agreement about the issues surveyed, justify scepticism about their value to measure the actual phenomenon of corruption. 10 On the other hand, a few surveys on experience with bribery were conducted both in particular countries and encompassing groups of countries. The best known of the latter is the International Crime Victim Survey (ICVS), conducted by the United Nations Interregional Crime & Justice Research Institute, which includes one variable related to corruption. Because of the scarcity of data, few studies on the relationship between perceptions and experience with corruption across assemblages of countries were conducted. Recently, Mocan (2004) studied data from ICVS and compared them with four perceptions indices, collected along different periods. His conclusion is that the perception of corruption in a country is mainly influenced by the quality of its institutions (proxyed by the risk of expropriation), and that, when this factor is compensated for, actual experience has no impact on the level of perceived corruption. 8 KK also uses surveys among the general population. 9 Those indices hold very high pairwase correlations which is only to be expected, as often they are based on similar or even identical samples. 10 See e.g. Johnston (2000), Søreide (2003) and (2006). 8

10 The Global Corruption Barometer presents a valuable opportunity to compare perceptions and experience within the same sample s. The importance of establishing the relationship, if any, between experience and opinions cannot be minimized, as reporting instances of bribery provides a presumably objective assessment of the actual incidence of corruption upon populations. Comparing experience with perceptions within the same sample allows one to investigate how the former relates to the latter. Let us start with the averages. Table 1 shows the averages of the variables belonging to the Effects group, plus Experience (base percentages weighted). Table 2 has the corresponding numbers for the variables of the Institutions group. Table 1: Average percentages, Effects and Experience Petty Grand Life Business Politics Perspective Experience All 73.3% 79.0% 41.3% 66.0% 69.9% 40.3% 11.5% Top % 61.7% 24.0% 56.7% 63.6% 42.2% 1.6% Bottom % 89.1% 51.3% 71.3% 73.5% 39.1% 17.3% It is apparent from Table 1 that Experience is more than eleven times as frequent in the Bottom group as in the Top (although the average frequency of Experience in the Top group is less than the margin of error of 4.4 percent points affecting the survey). Bottom group countries hold more pe s- simistic opinions about the overall effects of corruption than the Top, with the exception of Perspe c- tive Perspective systematically behaves at odds with the other variables, and in the sequel will be treated separately. Table 2: Average percentages, Institutions Customs Education Judiciary Health Police Parties Parliament Civil Registry Utilities Taxes Private Sector Media Military NGOs Religions All 44.9% 32.9% 49.7% 40.6% 52.3% 60.7% 51.1% 33.4% 31.3% 42.3% 41.8% 35.2% 26.2% 22.3% 23.4% Top % 15.0% 27.5% 21.7% 24.3% 47.7% 34.3% 13.9% 19.8% 26.4% 31.0% 32.6% 15.3% 14.6% 22.1% Bottom % 43.2% 62.5% 51.6% 68.6% 68.2% 60.9% 44.6% 38.0% 51.6% 48.1% 36.7% 32.5% 26.7% 24.1% Harsher evaluations among Bottom countries are also the norm for the variables from the group of Institutions (Table 2), with two exceptions, Media and Religions. It is also interesting to observe a few of the evaluations some of the Institutions got. The Police, with which people have 9

11 frequent contact, is among the worse-evaluated. However, so is Customs, 11 an institution remote from common citizens day-to-day experience. Other institutions that in principle are better known by citizens Education, Health, Civil Registry are midway. And certain others, that most citizens view at a considerable distance (Judiciary, Political Parties and Parliament) are among those that receive higher marks for perceived corruption. Finally, it is noteworthy that the military did not get bad marks in the group of poor countries, many of them with histories of military rule. Percentages alone do not explain a great deal. We are concerned with investigating the relationships between the sets of answers, in order to ascertain their mutual coherence. We first observe the marked differences between the Top and Bottom subsets. While among the Top countries low levels of experience go hand in hand with widely variable perceptions, among the Bottom very high perceived levels of corruption are associated with experiences distributed across a wide range. This results from the distribution of experience according to GDP-PC. While the GDP-PC vs. Experience correlation across the whole set of 60 countries is , this information is misleading, as the correlations are small in the income subsets, being (observe that it is positive) among the Top countries and only among the Bottom ones (Graph 1 and Table 3, where Greece and South Korea are enhanced; these two countries were not included in the calculation of the R 2 of 0.15 for the Bottom subset informed in the graph). Thus, the most one can say about the experience of bribery in what regards income is that it is fairly low among richer countries, and that experiences vary widely for those below GDP-PC of US$ 10,000. The disparate behaviour of Experience according to the income bracket justifies our systematically treating these subsets separately. 11 Apropos: Of the total population in Latvia, each year no more than 5-10% seek help from judicial institutions. However, 99% of the population, even though it does not even know where the Temple of Themis is located, has its own more or less justified view about the courts high level of corruption. Only every third person living in Latvia travels abroad each year. However, regardless of whether someone does or does not go on shopping trips to the Vilnius market, or on vacations to Greece or the Canary Islands, or whether this someone can afford neither the absolute majority of the respondents claim that customs officials and border guards are corrupt (and not just a little bit corrupt). Andrejs Vilks, Corruption Perception Restraint, Latvian Daily Neatkariga Rita Avize, Nov 27, 2002, p

12 Graph 1: GDP per capita vs. Experience with bribery. Experience 60% 50% 40% 30% 20% 10% R 2 = 0,15 0% GDP-PC Greece & SKorea Top 22 Bottom Linear (Bottom) GDP per capita correlates better with the subjective variables, and in almost all cases better among Top countries than among Bottom countries (Table 3), the exception being Perspective. Business and Politics correlate positively with GDP-PC in the Bottom subset (the higher the income, the more pessimistic is the perceived effect of corruption), but in the case of Business the correlation it is not significant. Table 3: Effects, Experience and GDP-PC All Top 22 Bottom 38 Petty ** Grand ** * Life ** ** Business ** Politics ** * 0.335* Perspective Experience ** * ** Significant (2-tailed) at the 0.01 level. * At the 0.05 level. The correlations of GDP-PC with the institutional variables also vary considerably between the income subsets. The highest correlation for both subsets involves Customs (respe ctively and ); also, for this variable the difference in absolute value between the correlations in the Top and Bottom subs ets is one of the lowest, The weakest correlations of GDP-PC in the Top subset are with Religions, Media (both less then -0.1) and Private Sector (-0.225), while in the Bottom subset they are Media, Relig ions and Health (all positive). With few exceptions, the other var i- 11

13 ables also exhibit low correlations with GDP-PC. Thus, in general, as far as the predictive power of income for opinions goes, it is stronger for perceived Effects of corruption than for Institutions and stronger among countries that are richer and less objectively affected by bribery than among poorer and more prone to bribery ones. The correlations with variables of the Institutions group are gene r- ally weaker, but the disparity between the income groups remains. One of our principal aims here is to ascertain how pragmatic experience as reported by respondents relate to their opinions. We will then begin with the relationships of the Experience variable with those concerning opinions about the Effects of corruption (Table 4). Table 4: Correlations Experience vs. Effects. Petty Grand Life Business Politics Perspective All 0.531** 0.433** 0.433** 0.308* Top * Bottom ** Significant (2-tailed) at the 0.01 level. * At the 0.05 level. At least half the relationships are not very close in the entire set to begin with, and all fall when calculated among countries belonging to the income subsets. Moreover, most of the correlations are negative and non-significant in the Top subset and especially small, and none is significant, in the Bottom subset. In words, for the Top, this means that, although discreetly, the less respondents reported experiences, the more pessimistic they manifested themselves about the various themes and conversely, and for the Bottom, that Experience is basically neutral: Experience does not inform opinions one way or the other. The same happens with almost all opinions about Institutions (Table 5), with one peculiar exception: Customs, again. This is the variable that presents the highest correlation with Experience in the Bottom subset, being significant at the 0.01 level. Most other correlations in this subset are very small and non-significant, including those with institutions that common citizens more likely have contacts. Even for the Police the correlation with Experience (significant at the 0.05 level) is only in the Bottom subset, being the highest negative in the Top, with (significant at the 0,05 level) the more respondents reported having had paid bribes to someone, somewhere, the better they evaluated the P olice. 12

14 Table 5: Correlations Experience vs. Institutions. Customs Education Judiciary Health Police Parties Parliament Civil Registry Utilities Taxes Private Sector Media Military NGOs Religions All 0.728** 0.469** 0.592** 0.463** 0.646** 0.297* 0.368** 0.511** 0.331** 0.506** 0.441** ** Top * * Bottom ** * ** Significant (2-tailed) at the 0.01 level. * At the 0.05 level. The disparate behaviour of the correlations can be visualized in graphical form in the examples of Experience vs. Petty and Experience vs. Health (Graph 2). Petty, together with Grand, is the be t- ter-behaved graph. Those corresponding to most of the other variables, both from the Effects group and from Institutions, are similar, usually being more scattered. Others (like Media) are markedly scattered about, as implied by the very low correlations (which in this case, as in others, are negative for both income groups ). Graph 2: How much Experience explains Petty and Health Petty 120% 100% R 2 = 0,07 80% 60% 40% 20% R 2 = 0,10 0% 0% 10% 20% 30% 40% 50% 60% Experience Top 22 Bottom 38 Linear (Bottom 38) Linear (Top 22) 13

15 Health 90% 80% 70% Graph 2: How much Experience explains Petty and Health 60% 50% R 2 = 0,01 40% 30% 20% 10% 0% 0% 10% 20% 30% 40% 50% 60% Experience Top 22 Bottom 38 Linear (Bottom 38) Such weak relationships between Experience and the opinion variables are counter-intuitive, notably concerning variables such as Petty, Life, Health, Education, Police etc. They presumably refer to institutions and to effects of corruption that would be more present and more easily within the grasp of the general population, more so than, say, the Grand or Business variables. One would perhaps not be surprised to find lower correlations between Experience and the latter two (and, maybe, Politics, Judiciary, Parliament etc.) than with Petty or Health. This is not verified, the correlations being similarly low (or negative) across the board. This means that comparing reported experiences with bribery across countries give scant information, if at all, about comparative assessments of the integrity of Institutions or Effects of corruption. Conversely, comparisons between such opinions give no clue about comparative levels of actual corruption. Indeed, if we control Experience vs. the opinion variables for GDP per capita, we find that no correlation is significant with Effects and that, among Institutions, only Customs (0.484 at the 0.01 level) and Police (0.315 at the 0.05 level) are significant. In contrast with the correlations involving Experience, those holding between the subjective variables are considerably higher. Not only that, they also vary much less across the income subsets. Table 6 describes the relationships within the Effects group, excluding Perspective. 14

16 Table 6: Correlations between perceptions about Effects. Petty Grand Life Business Top 22 Grand 0.949** Life 0.494* 0.549** Business * 0.583** Politics 0.625** 0.748** 0.670** 0.721** Bottom 38 Grand 0.847** Life 0.682** 0.617** Business 0.496** 0.479** 0.407* Politics 0.404* 0.530** ** ** Significant (2-tailed) at the 0.01 level. * At the 0.05 level. Therefore, learning the opinion of people about Petty corruption affecting countries gives a very good idea about what would be their opinions about Grand corruption. On the other hand, although the correlations between Life and Petty are not small, one would perhaps expect an even stronger relationship between them. At this point, a pertinent question concerning opinions is how evaluations of Effects of corruption relate to assessments of Institutions. It would be natural to expect that certain perceived Effects would be more linked with some Institutions than with others. So, it would not be remarkable if perceived problems with Civil Registry, Health, Education etc. reflected on opinions about Petty corruption, say. Likewise, if Parliament and Polit ical Parties are seen to be affected by corruption, then it would be natural to expect these variables to be strongly related to the perceived effects of corruption on the political life. An examination of the correlations shows that some of those direct expected relationships do appear, but not indiscriminately across the income subsets. Conversely, certain correlations that prima facie would not be expected to be strong prove otherwise. It is the case of Customs, correlating well with Petty and Life in the Top subset. Petty also correlates reasonably well with Customs in the Bottom subset. As an opinion about Effects might be informed by more than one single opinion about Instit u- tions, we will pursue the matter of how much the latter explains the former by performing multiple regressions on them. We will address each Effects variable in turn, dropping Religions, NGOs, Media and the Military, because each of them correlates weakly with the Effects variables. We will use the following terminology: Expected will refer to a set of explanations that one would reasonably expect for any given effect, given its nature. For instance, while it would be natural to expect that corruption in Education and in Health impact on the perceptions about the extent of over-the- 15

17 counter bribery and also on the significance of corruption in day-to-day life (and so these variables, among others, would comprise the Expected set of explanations for both Petty and Life), it would not be expected that corruption in the health system would have a sizable impact on the political life, say. The regressions with the Expected sets will be performed with and without the addition of the variable Experience. 12 Table 7 summarises the results. Petty Grand Life Business Politics Table 7: Effects explained by Institutions. All Top Bottom Independent variables adjusted R 2 S.E. adjusted R 2 S.E. adjusted R 2 S.E. All Expected (**) , All Expected (**) All Expected (**) All Expected (**) All Expected (**) In smaller type, the adjusted R 2 for the regressions including Experience. (**) See the text for the composition of these sets for each Effects variable. The regression for Petty using all variables over the whole set of countries leads to a multiple adjusted R 2 of 0.811, which is a good fit, but the standard error is big, Taking just the variables that in principle would have more to do with petty corruption (Education, Health, Police, Civil Registry and Taxes) and running the regression over this Expected set, we get an adjusted R 2 = and standard error of Within these limits, we can say that taking the whole set of countries, these variables reasonably explain the evaluations about Petty (and Grand) corruption. However, the picture changes when we focus on the income subsets. Whereas evaluations of institutions reasonably explain opinions about the extent of Petty corruption if we keep to richer countries, the explanatory power of such institutional assessments falls dramatically for the set of poorer coun- 12 Although the standard procedure is to find the least ensemble of independent variables that explains the dependent variable (thus getting parsimonious explanations), we will not pursue this path. 16

18 tries. When Experience is added to the regression, the fit slightly deteriorates. In fact, controlling Petty vs. Experience in the income subsets for the variables from Petty s Expected set, the correlations totally cease to be significant (respectively and 0.084). Evaluations about Grand corruption would have as natural explanations assessments of instit u- tions like Parliament, Parties and Judiciary, so we will use these to form our Expected set for this variable. We find that the explanatory power is very good within Top countries, falling for Bottom ones. This is one of the two exceptions where adding the Experience variable to the regression improves the fit, but only for the Top group. Adding Private Sector to the Expected set betters the fit when the regression is performed over the whole set and for the Bottom countries, but deteriorates it in the Top subset Taking Life, the explanatory power fall for both sets of independent variables. Expected is the same as was used for Petty. The differences of predictive power across the income divide are less than for Petty, but standard errors are bigger. Here Experience adds to the goodness of fit in the Top group. Controlling Life vs. Experience in the income subsets for the variables from its Expected set, the correlations become respectively and The situation for the effects of corruption on Business is much worse than for the previous variables. For the Expected set we choose Taxes, Utilities, Customs, Parties, Parliament, Private Sector and Judiciary. Table 7 shows that the opinions about the effects of corruption on Business are not very well explained in the Top subset and remain unexplained in the Bottom subset. Experience deteriorates the fits. Limiting the Expected set to fewer independent variables results in even worse outcomes. Lastly, the effects of corruption on Politics. It would be reasonable to find explanations in Parliament, Parties and Judiciary. The best fits involve all variables. The fit is moderately good with the Expected set for Top countries but not so for Bottom ones. Adding Private Sector betters the fit in the income subsets, while Experience deteriorates it. So, in summary: Petty is indeed acceptably explained by those perceptions of the incidence of bribery in institutions one would associate with it, but only in the set of richer countries. Among poor countries, the explanatory power is irrelevant however set of independent variables one chooses. 17

19 Grand is exceptionally well explained by the chosen Institutions variables in the Top group, such predictive power sharply falling among the Bottom countries, but still maintaining an arguable connection, with low standard errors. Estimating Grand for Top countries by the variables of its Expected group leads to a very good fit (adjusted R 2 = 0.914, standard error = 0.073, as shown in Graph 3). Grand (Top only) 120% Graph 3: How much Grand in Top countries is explained by its Expected set of Institutions. 100% 80% 60% 40% adjusted R2 = % 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Estimated Grand Life does not work especially well neither in the Top nor in the Bottom subset (adjusted R 2 of respectively about 0.5 and 0.4 for the Expected set of explaining variables, with a standard error of for the latter case). The opinions about the impact of corruption on Business remain unexplained for the Bottom subset and moderately explained in the Top one by the Expected set. Politics shows a slightly worse result within the Top for the Expected set of explanations, such deterioration being considerably more acute among Bottom countries. With the exception of Grand and Life for the Top subset, the addition of Experience deteriorates all regressions. 18

20 The unusual behaviour of perspectives about the future Every analysis performed over the variables led to distinctly peculiar results concerning the Perspectives variable. Around 40% of respondents in both income subsets consider that the future is bleak concerning the evolution of corruption, but the relationships of such opinion with the other perceptual evaluations of Effects of corruption are very weak (Table 8) and in fact non-significant. The highest positive correlation is only (with Life, in the Bottom subset). The correlations with Experience are all very low. Table 8: Correlations of Effects with Perspective. Petty Grand Life Business Politics Experience All Top Bottom No correlation is significant. The picture concerning the opinions about Institutions is similar. In the Top subset, the correlations range from a min imum of (with Police) to a maximum of (with NGOs,) being significant (at the 0.05 level) only with NGOs. In the Bottom group, the minimum is (Private Sector, not significant) and the maximum is (Utilities, significant at the 0.01 level). So, Perspective does not show to be connected with the other variables in any coherent way. The situation of this variable is worse than the others, because for those at least some higher significant correlations are found. Perspective seems to float by itself. This can be visually seen in the examples of dispersions (Graph 4) vis à vis Experience and Life. The graphs are very much dispersed. Graph 4: Examples of dispersions involving Perspective. Pespective 100% Perspective 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% 20% 40% 60% 80% 100% Life 0% 0% 10% 20% 30% 40% 50% 60% Experience 19

21 From the discussion so far it can be concluded that the variation of experience across countries is not a good predictor of variations of perceptions, irrespective of the income bracket considered. As the correlations run both ways, this means that comparing perceptions of corruption across countries does not furnish a reliable compass to assess comparative levels of bribery affecting common citizens. In contrast, opinions show a much better-behaved pattern, with exceptions. Some, but not all, opinions about the effects of corruption are reasonably explained by assessments of the inc i- dence of corruption in institutions among countries belonging to the upper income tier. For poorer countries, the predictive power of these assessments fall considerably. Adding experience to the explanations of perceived effects of corruption by assessments of the integrity of institutions actually deteriorates the explanatory power of such assessments. Outlook evaluations seem not to be significantly connected either with experience or with other opinions. Intra-country relationships between variables Weak or strong relationships between country averages give no information about the linkages among the same variables within each country. Thus, for example, from the generally low degree of relationship between Experience and opinions one cannot conclude that personal/household exper i- ence has little connection with opinions in any given country, but only that, if experience informs opinions, it does it differently across countries. It could be that, staying within each country, one would find stronger links between the variables in question. If we were to confirm this, then we would become endowed with country-specific numerical factors that, applied to each country, would permit to normalize results in order to allow comparing opinions across countries. In order to test whether or not this happens, we submitted the survey s country data to tests of dependence between variables. To do that, the crosstab statistics for each pair of variables was studied. Given a country and given two variables (say, Petty and Experience), we are interested in ascertaining whether or not respondents that reported having paid bribes are significantly more likely to hold a pessimistic opinion about the extent with which petty corruption is a problem than otherwise. The most common way to do that is by means of the χ 2 test. Here we slightly depart from the usual path and directly explore the hypergeometric distribution, characteristic of sampling procedures. 20

22 This allows for a better discrimination of cases than the χ 2 procedure. 13 We compare the frequency of the event Petty in the overall sample with the same event in the subsample defined by those who reported having had experience. If the frequency of the event Petty in the subsample is significantly higher (or lower) than the frequency in the overall sample, then we conclude that the two events are dependent. Thus, the outcome of a test pivots on how we define the level of signif icance, that is, the range of frequencies that establishes whether we are willing to accept the hypothesis that the events are dependent. The margin is: f ( 1 f )( n r) ε( r) = λπ, where nr n is the length of the sequence (the size of each country s sample); f is the frequency of the studied event in the sequence (say, the percentage of respondents that considered Petty corruption to be problematic); 14 r is the size of the subsample (say, the number of respondents that reported experience); λ π is the parameter of the elected level of confidence, corresponding to π = {1 ϕ( λ )}, 2 π where ϕ is the normal distribution function. We want to be as accommodating as possible concerning the rigour of the tests we want to apply, in order not to be guilty of bias towards rejection. Accordingly, for these tests we used a level of confidence of 90% (we accepted as dependent as many as 10% of all possible outcomes), corresponding to λ π = To apply a test, the absolute value δ of the distance between the sampled frequency and the frequency of the event in the original sequence is compared with ε(r). If δ > ε(r), the sequence is dependent relative to the test. For instance, take the variables Health and Experience in Argentina. The numbers are: n = 1005; f = 0.406; r = 71; and therefore ε(71) = Now we compute the frequency of the event Health among the 71 respondents that reported having had Experience. It is 43/71 = 0.606, and therefore δ = = Since δ > ε(r), we conclude that there is dependence between the two variables at the chosen level of confidence of 90%. 13 In fact, the χ 2 test leads to more situations where randomness cannot be rejected than the procedure used here. 14 Observe that this is not the weighted frequency for the variable, but the actual frequency of the event in the sequence. 21

23 Our intention was to ascertain whether or not the perceptions variables connected with corruption are dependent on the Experience variable and also among themselves. The rationale for applying the procedure was that, if experience informs perceptions, then those who have had household contacts with bribery would be more likely to hold pessimistic views than those who haven t. Similarly, we compared each of the variables of the Effects group with their respective sets of Expected variables (see the previous section) selected from the Institutions group. The tests could produce three types of outcomes: Randomness. If for country A Experience is random relative to Petty (say), then one cannot say that it is likely that actual experience with bribery significantly informed the opinions of the persons pertaining to that country s sample. Dependence by excess. When the frequency of the event under scrutiny (e.g. persons saying that corruption constitutes a problem in life) within the subset of respondents that have had experience with bribery is significantly higher than the frequency in the entire sample. Dependence by excess is what we are looking for. Dependence by deficiency, or lack. When the frequency of the event under scrutiny within the subset of respondents that have had experience with bribery is significantly lower than the frequency in the entire sample. We begin with the relationship of the Experience variable with the Effects group. The outcomes of the tests are summarised in Table 9. Table 9: Summary of tests for dependence between Experience and the Effects variables Petty Grand Life Business Politics Perspective Lack Random Dependent It turns out that in only 15 out of 60 countries respondents that have had experience with bribery were significantly more likely to answer that Petty corruption is a problem in their country than the country s overall sample. There were five countries (Guatemala, India, Indonesia, the Philippines and USA) where for these respondents it was significantly less likely that they would consider 22

24 petty corruption a problem than otherwise. For no less than 40 countries, the answers about experience and Petty corruption were relatively random. The panorama for the Grand variable is essentially the same, with only 12 countries exhibiting dependence between the variables. For seven there is non-randomness by lack of sufficient coincident answers (Germany, Guatemala, India, South Korea, Philippines, Portugal and USA), that is, respondents that experienced bribery were less likely to consider Grand corruption a problem in their countries than the incidence in the respective overall samples. The better-behaved instance concerns the variable Life, but even then only 25 countries showed dependence with Experie nce. On average, 72% of the relationships of the subjective Effects variables with Experience are either random of deficient, only 28% being dependent. In what regards the comparative picture between Top and Bottom countries, a partial summary is presented in Table 10, including only the relationships with Experience that proved to be dependent. Different behaviours appear, some of them more marked than others. Thus, Life shows to be dependent with Experience in a higher proportion of Bottom countries (47%) than of Top ones (32%). Something similar happens with Business (39% vs. 14%) and Perspective (37% vs. 14%). Even then, dependences with Experience remain fairly low within the groups. Table 10: Dependences between Effects and Experience by income group. All Top 22 Bottom 38 Petty 15 25% 5 23% 10 26% Grand 12 20% 4 18% 8 21% Life 25 42% 7 32% 18 47% Business 18 30% 3 14% 15 39% Politics 14 23% 4 18% 10 26% Perspective 17 28% 3 14% 14 37% In just two countries (Netherlands and Romania) dependence was found between Experience and all Effects variables (excepting Perspective). Taking just Petty, Grand and Life, in addition to Netherlands and Romania just four other countries (Estonia, Venezuela, Bulgaria and the Czech Republic) showed simultaneous dependences. Taking Grand, Business and Politics, only three satisfied at the same time the dependence criterion with Experience: The already seen Netherlands and 23

25 Romania, plus Denmark. 15 Conversely, in 16 countries all relations of Effects (excepting Perspective) with Experience are random: Bosnia/Herzegovina, Brazil, Canada, Ecuador, France, Hong Kong, Ireland, Israel, Japan, Luxembourg, Malaysia, Peru, Poland, Switzerland, Taiwan and Uruguay. Adding Perspective, the number is 13. In the previous section we have seen that the correlations between the Effects variables and Experience were low. Perhaps we could find better relationships if we limited the countries to those that exhibited dependence between Experience and each Effects variable in turn. Table 11 shows that the correlations obtained are even worse than previously. Only Grand, with among 12 countries, escapes from generally negative or near zero correlations. Table 11: Correlations between Effects and Experience among countries where there is dependence between the variables. R # of countries Petty Grand Life Business Politics Perspective What all this means is that the conclusions of the previous section are confirmed for the variables of the Effects group, now within countries: Personal or household experience with bribery does not consistently inform people s opinions about the effects of corruption, either about Grand and Business (which perhaps would be expected) or about Petty or Life (which, presumably, would be linked with personal histories), or about Politics. There are differences between richer and poorer countries, but they remain within the picture arising from the overall numbers. As before, the re a- soning is symmetrical for the variables, and therefore assessing people s opinions about the effects of corruption does not furnish a reliable indirect measure of actual corruption happening in the majority of countries. Indeed, for a number of countries, all opinion/perceptual variables are relatively 15 Even at the 70% confidence level (accepting as dependent 30% of all possible outcomes of the tests), the distribution of dependences of Experience vs. the subjective variables is: Petty 24 countries; Grand 31; Life 34; Business 28; Politics 26; Perspective 25. However, there is a trade-off, as dependences by deficiency also grow: Petty 9; Grand 9; Life 4; Business 5; Politics 7; Perspective 6. With this, as much as 56% of the relationships turns out to be non -positively-dependent even at that more than permissive level of confidence. 24